The investment factor: why companies that spend big often deliver small returns

Companies that invest aggressively in growth tend to deliver lower returns than their more cautious peers. This counterintuitive pattern, known as the investment factor, has been documented across six decades of data. And it has profound implications for how we should think about Big Tech's extraordinary spending binge.

I've been watching the numbers with a mix of fascination and disbelief. Over the past two years, the so-called hyperscalers have embarked on a capital-spending binge with little modern precedent. Microsoft, Amazon, Alphabet, Meta. In 2024 alone, these four companies spent well over $300 billion on capital expenditure, much of it AI-related. Estimates for 2025 run higher still, around $340–360 billion. That's spending on a scale that rivals entire national industries.

Companies that once prided themselves on being asset-light are now pouring money into data centres, specialist chips and power infrastructure. Back in 2018, the same four companies spent roughly $70–80 billion combined. We're looking at a four to fivefold increase in seven years.

The natural assumption is that aggressive investment generates extraordinary returns. More spending, more growth, more profit.

But academic finance has spent decades studying companies that invest heavily versus those that don't. The findings run directly counter to what most investors expect.

This is a story about what happens when ambition meets arithmetic.

The assumption everyone makes

We're taught from the start that reinvestment drives growth. Companies plough profits back into the business, expand capacity, hire talent, build new products. The flywheel spins faster. Shareholders get rich.

It's an intuitive story. And it shapes how most people think about investing.

Active fund managers chase these narratives. They hunt for companies with ambitious expansion plans, rising capital expenditure, bold bets on the future. The financial media celebrates them. Analysts reward them with "buy" ratings.

Growth stocks, almost by definition, are companies that reinvest aggressively rather than returning cash to shareholders. That's supposed to be the whole point.

But what if this assumption is wrong? What if the relationship between corporate investment and shareholder returns isn't positive at all?

What 60 years of data actually show

Companies that invest conservatively outperform those that invest aggressively. The pattern holds across decades, different markets, and multiple research methodologies.

In 2015, Eugene Fama and Kenneth French expanded their famous three-factor model to include two new variables: profitability and investment. The investment factor, which they called CMA (Conservative Minus Aggressive), measures asset growth. Companies that grow their asset base slowly go into one bucket. Companies that expand rapidly go into another.

The conservative bucket wins. Consistently.

Fama and French found something else. When they added profitability and investment to their model, the traditional value factor became largely redundant. The thing we thought was "value" was partly explained by investment behaviour all along (Fama & French, 2015).

The numbers are stark. According to data from Ken French's research library, low-investment small-cap stocks outperformed their high-investment counterparts by 5.39% annually between 1964 and 2023. That's not a marginal edge. It's a chasm.

Why would this be true? The valuation logic is straightforward. A company's market value equals its expected future earnings, discounted back to today, minus the investment required to generate those earnings. All else equal, a company pouring money into new projects is worth less right now. Those projects consume cash that would otherwise flow to shareholders.

This isn't some behavioural quirk. It's arithmetic.

Recent research from Verdad Capital puts flesh on these bones. Looking at data from 1997 to 2025, they found that companies in the highest quintile for capital expenditure generated 5.2% annualised EBIT growth over three years. Impressive. Companies with high free cash flow managed 1.6%.

Companies that spend aggressively on capital expenditure generate faster earnings growth. But, as we'll see, that doesn't translate into better returns. Source: Verdad Capital (2025)

But when Verdad looked at actual stock returns rather than earnings growth, the relationship flipped completely. High capex didn't just fail to deliver a premium. It delivered a penalty. Across every time horizon Verdad measured, the most aggressive spenders underperformed the most conservative by 0.9% to 1.4% annually.

The punchline: high capex doesn't just fail to deliver a return premium; it delivers a penalty. The most aggressive spenders underperformed the most conservative by 0.9% to 1.4% annually across every time horizon measured. Source: Verdad Capital (2025)

Why does this happen?

Two explanations. Both probably contain some truth.

The first is rational. Companies invest heavily when their cost of capital is low. A low cost of capital means the market isn't demanding high returns to compensate for risk. So when a company ramps up spending, it's signalling something about its own risk profile. Lower risk, lower expected returns. Nobody's being fooled.

This explanation is neat. Perhaps too neat.

The second explanation is behavioural, and messier. Managers like building empires. They prefer running large organisations to small ones, even when shareholders would be better served by restraint. Executive compensation often rewards revenue growth and asset expansion.

There's also overconfidence. Management teams consistently overestimate the returns their projects will generate. They see opportunities everywhere. They believe their competitive advantages will persist. They discount the possibility that the new factory, the acquisition, the expansion into adjacent markets might destroy value.

And defensive spending plays a role too. Once one company in an industry starts investing heavily, others feel compelled to follow. Nobody wants to be left behind. The result is an arms race that benefits nobody except suppliers of capital equipment.

The Verdad research points to a specific mechanism. High capital expenditure correlates strongly with low return on assets. Companies in the highest capex quintile had gross profit margins 11.6 percentage points lower than those in the lowest quintile. Heavy spenders are systematically less efficient with their capital.

The mechanism: companies with high capital expenditure are systematically less efficient with their assets. Heavy spenders had gross profit margins 11.6 percentage points lower than conservative spenders. Source: Verdad Capital (2025)

And return on assets, unlike capex, does predict stock returns. Across every time horizon, companies with higher ROA delivered better shareholder returns.

Unlike capex, return on assets does predict stock returns. The logic chain is complete: aggressive spending correlates with lower efficiency, and lower efficiency correlates with worse returns. Source: Verdad Capital (2025)

The logic chain is tight: aggressive spending correlates with lower efficiency, and lower efficiency correlates with worse returns. Shareholders fund expansion that dilutes their ownership without delivering commensurate value.

The caveat

The picture isn't quite as clean as I've made it sound.

Research from Cooper, Gulen and Ion (2024) complicates matters. They dug into what's actually driving the investment factor's predictive power. And they found something unexpected.

The returns aren't coming from where theory suggests they should.

When you measure investment using traditional capital expenditure or property, plant and equipment, the factor doesn't perform nearly as well. The real predictive power comes from changes in inventory and accounts receivable. Working capital.

This is awkward. The theoretical story about companies consuming cash through expansion doesn't map neatly onto what's generating the returns in the data. As Larry Swedroe summarised it: "the empirical success of the models is not linked to financial theory."

So we have a pattern that works. We have a theory that sounds plausible. But the two don't connect in the way we'd expect.

Does this invalidate the finding? Not necessarily. The return premium is real. That's documented beyond reasonable doubt.

But we should be honest about what we don't fully understand. The mechanism remains contested.

Intellectual humility matters here. The investment factor works. We're not entirely sure why.

Big Tech as a test case

Which brings us back to the hyperscalers.

For most of their existence, Microsoft, Amazon, Alphabet and Meta operated as relatively asset-light businesses. Software scales beautifully. Write the code once, distribute it infinitely, collect the margins. That model made them some of the most profitable companies in history.

The AI era is changing the equation.

Data centres require physical infrastructure. Servers. Cooling systems. Vast amounts of electricity. Specialist chips that cost tens of thousands of pounds each and need replacing every few years. The capital requirements are enormous and ongoing.

These companies are transforming from software businesses into something closer to industrial operators. The comparison isn't perfect, but there are echoes of the utilities sector: heavy upfront investment, long asset lives, returns that depend on capacity utilisation.

Verdad's Brian Chingono calls the current AI spending binge a "gamble." That's not hyperbole. The returns on this infrastructure are uncertain. Will enterprises pay enough for AI services to justify the investment? Will the technology evolve so quickly that today's data centres become obsolete before they've paid for themselves?

What the investment factor tells us: history suggests caution.

Companies that dramatically ramp up capital expenditure tend to earn lower returns for shareholders than those that don't. It doesn't mean these specific companies will underperform. It means the odds aren't automatically in their favour because they're spending big.

The market may already be pricing this in. Or it may not.

What the investment factor means for investors

This isn't a case for avoiding growth companies. Some aggressive investors will generate spectacular returns. The point is that you can't reliably identify which ones in advance.

Active fund managers spend their careers chasing exactly these stories. The next Amazon. The company whose bold investment programme will reshape an industry. They visit headquarters, interview management teams, build elaborate financial models. And yet the evidence on their collective performance is brutal. Most underperform a simple index fund over any meaningful time horizon.

The investment factor helps explain why.

When managers get excited about a company's growth prospects, that excitement is usually already in the share price. The market knows about the new factory, the acquisition strategy, the ambitious expansion plans. It's priced in. What remains is execution risk, competitive pressure, and the tendency for large investments to generate disappointing returns on capital.

Passive investors sidestep this entirely. A global index fund owns everything. Some companies will invest conservatively and do well. Others will spend aggressively and struggle. You capture the market return without betting on your ability to distinguish winners from losers.

Don't pay a premium for growth stories. The arithmetic rarely works in your favour.

The bottom line

The hyperscalers will keep spending. $300 billion this year, probably more next year. The financial press will continue celebrating their ambition. Analysts will build models projecting extraordinary returns from AI infrastructure.

Maybe they'll be right.

But 60 years of data suggest we should be sceptical of the assumption that aggressive investment translates into shareholder value. The relationship, if anything, runs the other way.

This doesn't mean Big Tech is doomed. It means the market has seen bold investment programmes before. Most disappoint.

For ordinary investors, the practical takeaway is liberating. You don't need to guess which growth stories will pay off. You don't need to evaluate management's capital allocation skills or forecast returns on AI infrastructure. A low-cost index fund captures whatever the market delivers, without the gamble.

Sometimes the smartest investment decision is the one you don't make.

Resources

Fama, E. F. & French, K. R. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1–22.

Cooper, M., Gulen, H. & Ion, M. (2024). The use of asset growth in empirical asset pricing models. Journal of Financial Economics, 151.

Chingono, B. (2025). Investing for growth. Verdad Capital Research.

Recently on TEBI

Is Terry Smith right to blame index funds for his struggles?

Buffer ETFs: protection, illusion or expensive reassurance?

Why do mutual funds underperform? Because most “skill” is actually luck

You understand the evidence. Now find an adviser who believes it too

Our Find an adviser directory lists professionals committed to low-cost, globally diversified portfolios and straight answers. Connect with someone who'll respect your knowledge instead of trying to sell you expensive active funds. Search your area now.

Stay connected: YouTube | LinkedIn | X

For advisers: Connect with Robin on LinkedIn